Professional Documents
Culture Documents
Past, Present, and Future of Basin and Petroleum System Modeling
Past, Present, and Future of Basin and Petroleum System Modeling
550 Past, Present, and Future of Basin and Petroleum System Modeling
Figure 1. The effect of seismic
technology on discovery and re-
covery efficiency (dashed line) is
shown throughout 50 yr for the
United States (after Weimer and
Slatt, 2006; used with permission
from AAPG). Gray line represents
a normalized curve. Gray arrows
indicate major technology de-
velopments, and white arrows
show timing of key concepts for
deep water settings. 3-D = three-
dimensional; 4-D = four-
dimensional.
AAPG Annual Conference and Exhibition contri- charge may increase the effect of BPSM over the next
butions from 2012 to 2016 (Figure 2B, C). In this decade. A growing BPSM community as evidenced by
study, seal accounted for 45% of well failures, and increasing archived AAPG abstracts containing “basin
charge accounted for 30%. However, these elements modeling” for the years between 1980 and 2010 is
represented only 2% and 14% of the research focus a step in the right direction. However, the current
(Figure 2B, C). In contrast, poor quality or missing misalignment of risk and research focus needs to be
reservoir accounted for 10% of the well failures recognized.
but received almost 70% of the research focus
(Figure 2B, C).
Taken together, these data show that the key risk HISTORICAL PERSPECTIVE
elements of seal and charge are not yet rigorously
addressed and that increasing research focus in these In the late 1970s, early computer programs were
areas could prove beneficial. Decreasing the number developed to quantify subsurface processes such as
of well failures because of improved assessment of burial history and source rock maturation through
552 Past, Present, and Future of Basin and Petroleum System Modeling
SURVEY Table 1. Question 1: What Was Your Main Motivation to Attend
This Conference?
Since the 1996 study (Marzi and Crowley, 2003), no
Response Response
comparable survey was acquired, and the frequency of
Answer Options Percent (%) Count
the application of basin modeling in exploration and
the overall complexity of industry projects remained (1) To share your work 39.6 21
unclear. To better understand how basin modeling (2) To see other people’s work 47.2 25
is used in exploration and appraisal, academia, re- (3) To network 13.2 7
search, and government organizations, a multiple- (4) To enjoy the beach 0.0 0
choice questionnaire was distributed to all participants
Responses show that most participants joined the AAPG Hedberg Research
of the AAPG Hedberg Research Conference on the Conference to see other people’s work. Of the participants, 53 answered this
Future of Basin and Petroleum System Modeling, question, and 0 skipped it.
held in Santa Barbara, California, April 3–8, 2016.
The questionnaire included 14 questions about how SURVEY RESULTS
people build basin models and how the results are
applied during exploration and appraisal decision- Question 1: Motivation for Participation
making. David Curry will present the results of
a second survey in a future issue, focusing on the Nearly half the respondents attended the Hedberg
most important future directions in BPSM (Curry, Research Conference to see the scientific work of
2018). That survey focuses on how to improve others (Table 1). The meeting consisted of oral and
BPSM and future research directions, whereas the poster sessions on rock properties and pressure pre-
questionnaire presented here focuses on a de- diction, case studies, fluid migration, geochemis-
scription of the current status of basin modeling try, geodynamics and heat flow, risk and uncertainty,
usage. new tools and workflows, and structural and strati-
The questionnaire for the study described in this graphic complexity. These themes grew naturally
paper was designed in multiple-choice format to from the abstracts submitted for consideration by the
make the answering process easy and to encourage conference conveners.
participation. Of the 98 registered participants, 53
answered the questionnaire. Some questions were
skipped by individuals, which resulted in 47 answers Question 2: Software Preference
for question 7. Most questions have more than 50
answers. The survey did not track background in- Most participants (56.6%) answered that they use
formation of the participants to maintain anonymity, PetroMod most commonly as a modeling tool, whereas
but every participant was assigned a random identi- 17% use Temis (Table 2). Thirteen percent use Trinity,
fication number so that it is possible to track the and another 13% use software like MigMOD, Migris,
answers that belong together (i.e., the same user Novva, TecMod, or proprietary software. Personal
identification). communications with colleagues across the industry
Although no direct association between the in-
dividuals who answered the survey and their pro- Table 2. Question 2: Which Basin Modeling Tool Do You Use
fessional affiliations can be made, we assume that the Most Frequently?
participants in the survey reflect the overall affiliation
Response Response
composition of conference participants: 62.2% came Answer Options Percent (%) Count
from the industry (exploration and production and
service), 27.6% were from academic institutions, and (1) Temis 17.0 9
10.2% were from government organizations. Many (2) Trinity 13.2 7
major and smaller oil companies were represented, as (3) PetroMod 56.6 30
were large and small software companies. (4) Other (please specify) 13.2 7
A select number of questions will be discussed in Responses show that PetroMod is dominant. Of the participants, 53 answered this
the next section. question, and 0 skipped it.
554 Past, Present, and Future of Basin and Petroleum System Modeling
Table 5. Question 5: How Do You Normally Calibrate Migration 4. Nature of seal influencing when and what column
to Match Available Fluid Data? height the prospect could hold (e.g., correlating
with phase, lithology, and burial history).
Response Response
Answer Options Percent (%) Count
If each uncertainty parameter were populated
(1) I change source rock depth, 8.3 4 with a base, low, and high case, we would end up
presence, and maturity. with X3 combinations where X is the number of
(2) As above, plus I change source 12.5 6 uncertainties. Using four uncertainty parameters
rock character (including kinetics). results in 43 or 64 possible scenarios for migration.
(3) As above, plus I change carrier 39.6 19 Most modelers feel comfortable to describe their
bed, faults, and other parameters petroleum system of investigation without running
affecting the plumbing system. all possible combinations. We may have good un-
(4) As above, plus I change 39.6 19 derstanding of the range of specific uncertainty
parameters specific to the parameters, but more commonly, modelers narrow
migration method I am using. the options and focus on a single explanation. For
Responses show that most modelers (almost 80%) assume that the basin model example an extreme scenario is that basalt could be
migration is strongly influenced by the plumbing system of the carrier beds and interpreted as carbonate, based on seismic veloci-
the migration algorithm. Of the participants, 48 answered this question, and 5
skipped it.
ties, which would have a dramatic effect on the
petroleum system and the charge model.
One reason why the preferred number of mi-
behavior. Accordingly, instead of changing parame-
gration models or scenarios does not exceed 15 might
ters unique to a particular migration method to cal-
be because of differences in the perceived meaning
ibrate a model (e.g., noise forced onto the IP method),
of “migration models” in the survey. Some survey
it may be more reasonable to test different migration
participants interpreted the term to represent fun-
methods with different physical driving mechanisms.
damentally different geological concepts based on
Overall, modelers use a wide variety of methods to
the discussion we had at the end of the conference.
calibrate migration, and there appears to be no
Assuming that each geological concept requires sev-
standardized workflow that the community as a
eral migration runs to be analyzed, this would increase
whole follows.
the number for migration models or scenarios.
556 Past, Present, and Future of Basin and Petroleum System Modeling
Table 9. Question 9: Do You Use Basin Modeling to Predict Pore Table 11. Question 11: Which Method Best Describes Your
Pressure? Approach toward Evaluating Source Rock Presence?
Question 13: Future Research Focus the most important topic for future research. Many
participants (15.1%) answered “other” for this ques-
This question about future research overlaps with
tion and supplemented their response with com-
Curry’s survey (Curry, 2018). Because the same
ments mainly associated with risk and uncertainty
community was asked for their opinion on future
analysis, thereby supporting answer 5. In contrast,
research, it is not surprising that the results are
thermal maturity prediction and source rock pres-
nearly identical. The most popular answer, at 28%
ence represented the lowest need for future research.
(answer 5) of survey respondents, suggested that the
This could be because of well-established workflows
need to improve workflows and risk evaluation
on those topics.
methods should be the primary focus of future re-
search (Table 13). Followed closely at 26.4% were
those who view migration and expulsion issues as Question 14: Most Important for Basin and
Petroleum System Modeling Community
Table 13. Question 13: Which of the Items below Should Be the
Primary Focus of Future Research in Basin and Petroleum System A lively discussion on the future of the BPSM dis-
Modeling? cipline concluded the meeting. Attendees voted
favorably for continuing to hold Hedberg Research
Response Response Conferences on an every-other-year basis (Table 14).
Answer Options Percent (%) Count
(1) Studies related to pressure, rock 18.9 10
mechanics, and facies population. PATTERNS IN THE SURVEY RESULTS
(2) Studies related to better maturity 3.8 2
modeling. We attempted to reveal associations among survey
(3) Studies related to better 7.5 4 responses using both descriptive and statistical
predicting source rock presence. methods. Statistical hierarchical clustering analysis
(4) Studies focusing on migration 26.4 14 revealed that no robust grouping could be established.
and expulsion-related uncertainties. Descriptive statistics were applied mainly focusing on
(5) Studies on improving general 28.3 15 the differences and similarities of different software
workflows and risk methods. user groups within the BPSM community. The fol-
(6) Other (please specify). 15.1 8 lowing statements can be made.
Responses show that most participants agree that hydrocarbon migration and
charge risking workflows need more focus in future research. Of the par- • Temis users are the group with the most people
ticipants, 53 answered the question, and 0 skipped it. running a high number of migration scenarios to
558 Past, Present, and Future of Basin and Petroleum System Modeling
determine charge risk (see Table 15: highest per- provided risking functionalities. The reason for
centage in question 6/answer 2 is in the Temis user this outcome might be related to the complexity of
group, whereas Trinity and PetroMod have highest the tools. The more complex the tool, the more
values in question 6/answer 1). Most Trinity and hesitant geologists might be to use it.
PetroMod users run not more than 15 different • Approximately 30% of all PetroMod users com-
migration models. This is a surprising result because monly do pore pressure prediction (question
a fast and simple migration algorithm in Trinity 9/answer 1) followed by Temis with 22.2%
enables the user to run and test more scenarios (question 9/answer 1) and Trinity with 14.3%
compared with full physics Temis or PetroMod (question 9/answer 1). Because PetroMod calcu-
solutions, which need more time. lates pore pressure by default and it is much more
• Opinions on how to deal with faults seem to be complex to integrate a pressure estimation into
most aligned within the PetroMod user group Trinity, the results are not surprising.
because over 80% of the participants indicated • All Trinity users (question 10/answer 4 = 100%)
that only in the case of hard evidence would and most Temis users (question 10/answer 4 =
faults be integrated into the model (question 55.5%) use various methods and data to predict
7/answer 3). Temis users concur with 67% (high temperatures. PetroMod is the only user group that
number in question 7/answer 3). Only the Trinity focuses mainly on a heat flow–based approach to
community is split into two parties: one using determine the thermal regime (question 10/answer
closed fault properties by default (question 2 = 45.5%).
7/answer 1) and the other party using fault prop- • Approximately 37.5% of Temis users use geo-
erties only in the event of hard evidence (question statistical methods and rock properties from seis-
7/answer 3). mic data to populate lithologies in a basin model
• In terms of calibrating migration, the PetroMod (question 12/answer 3), but only 18.5% of Pet-
user group answered question 5 by using all four roMod users (question 12/answer 3) and 0% of
possible answers (question 5/answers 1–4) with Trinity users do so, although all tools have the
most saying they would change the geological con- ability to do so.
cept to calibrate migration and would not change
parameters specific to the migration method. The The above statements are based on Table 15,
Temis user group answered this question by using which shows the underlying data expressed in per-
three possible answers (question 5/answers 2–4) with centage amounts for each answer, distinguished by
the majority agreeing with the PetroMod users. user group.
However, Trinity users selected only two different
answers (answers 3 and 4) with the majority (>70%)
of participants changing parameters specific to the RECENT AND POSSIBLE FUTURE
migration method. DEVELOPMENTS
• Only the PetroMod user group has a preferred
migration method, which is IP (highest in ques- In the past, basin models were used mainly as a one-
tion 4/answer 3 = 33.3%) followed by Darcy way technology, where the model predictions were
(question 4/answer 2 = 27.8%). However, most the end product and were not used as inputs for other
Trinity and Temis users state they use any mi- disciplines. However, this has changed over the last
gration method to calibrate their basin model two decades, as basin modeling has become more
to match fluid observations (highest in question integrated into an iterative, full-cycle workflow with
4/answer 4 for both). higher resolution. Rock properties from seismic data
• The Trinity users are the only group in which all are fed into basin models to simulate pore pressure
participants use the risking functionalities provided and other rock parameters (Kacewicz and Xu, 2006).
by the software. In other words, 37.5% of all Temis Derived basin model rock or fluid properties can now
users (question 3/answer 2), 27.3% of all PetroMod be used in seismic analysis and thus provide an op-
users (question 3/answer 2), and 0% of Trinity users portunity for an iterative approach. Reservoir quality
(question 3/answer 2) do not use the software- predictions use pressure and temperature histories
The question numbers are indicated by Q1–Q14. Numbers in first column represent the answers available for each question. The percentage levels are indicated by arrows pointing upward (12:00) for very high values, pointing toward
1:30 for high values, pointing toward 3:00 for intermediate values, pointing toward 4:30 for low values, and pointing downward (6:00) for very low values.
derived from basin models as input to the reservoir REFERENCES CITED
quality models, which are used to either predict
porosity or evaluate if the pressure–temperature Bjørlykke, K., K. Hoeg, J. I. Faleide, and J. Jahren, 2005, When
history can explain the measured porosities (Taylor do faults in sedimentary basins leak? Stress and deformation
in sedimentary basins; Examples from the North Sea and
et al., 2010). These are examples where basin model
Haltenbanken, Offshore Norway, a discussion: AAPG Bul-
predictions become important input parameters for letin, v. 89, no. 8, p. 1019–1031, doi:10.1306/04010504118.
other disciplines, which changes the traditional us- Curry, D. J., 2018, Future directions in basin and petroleum
ability of basin models. systems modeling: Building more useful models a sur-
In the future, the BPSM community hopes to vey of the community: AAPG Bulletin, doi:10.1306
/1208171615217152.
advance research on better risking methods. The Hantschel, T., and A. I. Kauerauf, 2009, Fundamentals of basin
principal physics behind migration is understood, and petroleum systems modeling: Berlin, Springer, 476 p.
but which mechanism is dominant on a basin scale Hunt, J. M., 1979, Petroleum geochemistry and geology: San
and which anomalies affect migration routes the Francisco, W.H. Freeman and Company, 617 p.
Kacewicz, M., and W. Xu, 2006, High-resolution prediction
most are not yet well understood. Any progress in
of rock properties and hydrocarbon charge through an in-
these fields (risking and migration) could have tegrated basin modeling/seismic inversion approach (abs.):
a major effect on the future of BPSM and explo- AAPG International Conference and Exhibition, Perth,
ration success rates. Western Australia, November 5–8, 2006, accessed April 14,
2017, http://www.searchanddiscovery.com/pdfz/abstracts
/pdf/2006/intl_perth/abstracts/ndx_kacewicz.pdf.html.
Karlsen, D. A., and J. E. Skeie, 2006, Petroleum migration,
CONCLUSIONS faults and overpressure, Part I: Calibrating basin mod-
elling using petroleum in traps—A review: Journal of
Comparing well failure analysis with current research Petroleum Geology, v. 29, p. 227–256, doi:10.1111
focus reveals that the principal elements responsible /j.1747-5457.2006.00227.x.
Marzi, R., and B. Crowley, 2003, Where did we come from? A
for dry holes do not receive the research focus they 1996–1997 hydrocarbon system technology benchmark
would need. For example, petroleum charge is a study, in S. J. Duppenbecker and R. W. Marzi, eds.,
largely underestimated risk when evaluating pros- Multidimensional basin modeling: AAPG/Datapages
pects, but the research attention charge receives is Discover Series 7, p. 1–8.
Rayeva, N., N. Kosnazarova, Z. Arykbayeva, and D. Shaikhina,
minimal. Therefore, a growing BPSM community can
2014, Petroleum systems modeling and exploration risk
have a significant effect in the future to mitigate assessment for the eastern margin of the Precaspian Basin:
charge related dry holes. SPE Annual Caspian Technical Conference and Exhibi-
Based on contributions, discussions, and the re- tion, Astana, Kazakhstan, November12–14, 2014, SPE-
sults of a survey executed during the AAPG Hedberg 172332-MS, 12 p.,doi:10.2118/172332-MS.
Schoener, R., V. Lueders, R. Ondrak, R. Gaupp, and P. Moeller,
Research Conference in Santa Barbara in April 2016,
2008, Fluid-rock interactions, in R. Littke, U. Bayer,
the following conclusions can be made. D. Gajewski, and S. Nelskamp, eds., Dynamics of com-
plex intracontinental basins. The Central European Basin
• Improved understanding of hydrocarbon migration System: Berlin, Springer-Verlag, p. 125–153.
Sweeney, J., and A. K. Burnham, 1990, Evaluation of a simple
is needed, as are more flexible risking workflows.
model of vitrinite reflectance based on chemical kinetics:
• Source rock maturity and temperature prediction is AAPG Bulletin, v. 74, no. 10, p. 1559–1570.
well understood with several different workflows Taylor, R. T., M. R. Giles, L. A. Hathon, T. N. Digg, N. R. Braunsdorf,
available, all leading to satisfactory results. G. V. Birbiglia, M. G. Kittridge, C. I. Macaulay, and
• Basin and petroleum system modeling is increasingly I. S. Espejo, 2010, Sandstone diagenesis and reservoir quality
prediction: Methods, myths, and reality: AAPG Bulletin,
used to predict pore pressure and reservoir quality at v. 94, no. 8, p. 1093–1132, doi:10.1306/04211009123.
field scale, which opens new opportunities to in- Tissot, B. P., and D. H. Welte, 1978, Petroleum formation and
tegrate BPSM workflows and concepts with other occurrence: Berlin, Springer-Verlag, 699 p., doi:10.1007
technologies, such as seismic rock property analysis /978-3-642-96446-6.
and reservoir quality modeling. Waples, D. W., 1980, Time and temperature in petroleum
formation: Application of Lopatin’s method to petro-
• Rock properties are populated in basin models leum exploration: AAPG Bulletin, v. 64, no. 6, p. 916–926.
mainly through GDE maps and not through seis- Weimer, P., and R. M. Slatt, 2006, Petroleum geology of deep-
mic or geostatistical methods. water settings: AAPG Studies in Geology 57, CD-ROM.